{"answers":{"affected":"3 open source packages ship an affected version, including keras. It affects 1 product.","exploited":"There is no known exploitation of CVE-2025-9906 in the wild and no public exploit code.","fix":"Upgrade keras to 3.11.0 or later.","what":"CVE-2025-9906 is a high-severity unsafe deserialization vulnerability in Keras-team Keras, published on 19 Sep 2025. It can only be exploited from the same machine or local network, and only if a user opens a file or follows a link."},"fix_prompt":"Check this codebase for CVE-2025-9906 and fix it if it is affected.\n\nWhat it is: CVE-2025-9906 is a high-severity unsafe deserialization vulnerability in Keras-team Keras, published on 19 Sep 2025. It can only be exploited from the same machine or local network, and only if a user opens a file or follows a link.\nExploitation: There is no known exploitation of CVE-2025-9906 in the wild and no public exploit code.\n\nAffected packages and fixed versions:\n- keras (PyPI): affected all versions; no fix listed\n- keras (PyPI): affected all versions; fixed in 3.11.0\n- keras (PyPI): affected all versions; fixed in 3.11.0\n\nTask:\n1. Search the dependency manifests and lockfiles (for example package.json, package-lock.json, requirements.txt, poetry.lock, uv.lock, pom.xml, build.gradle, go.mod, Cargo.lock, Gemfile.lock) and any Dockerfiles for the affected packages, including transitive dependencies.\n2. If an affected version is in use, upgrade it to the fixed version for its release line with the project's package manager, so the lockfile is regenerated. Prefer the smallest upgrade that fixes it.\n3. If there is no fix, or the upgrade would break the build, do not guess: explain the risk and the safest workaround instead.\n4. Run the tests and summarise what changed and why.\n5. If nothing in the project is affected, say so and change nothing.\n\nReference: https://insights.sbomify.com/CVE-2025-9906/ (Markdown: https://insights.sbomify.com/CVE-2025-9906/index.md)","id":"CVE-2025-9906","markdown":"https://insights.sbomify.com/CVE-2025-9906/index.md","name":"CVE-2025-9906","record":{"aliases":["AZL-67508","DEBIAN-CVE-2025-9906","GHSA-36fq-jgmw-4r9c","PYSEC-2025-76","UBUNTU-CVE-2025-9906"],"analysis":{"component_type":"library","fix_type":"patch","impacts":["code_execution"],"network_reachable":"adjacent_or_local","non_default_config":false,"product_category":"Keras deep learning library","requires_auth":"low_privilege","summary":"Attackers can execute arbitrary code by loading a malicious Keras model archive that disables safe mode and exploits Lambda layers.","user_interaction":"required","vuln_class":"deserialization"},"analysis_generated_by":"qwen3.8-flash-next","assigner":"Google","cpe":[{"product":"keras","vendor":"keras"}],"cvss":{"score":8.6,"severity":"high","source":"cna","vector":"CVSS:4.0/AV:L/AC:L/AT:N/PR:L/UI:P/VC:H/VI:H/VA:H/SC:H/SI:H/SA:H/AU:Y/R:A","version":"4.0"},"cwes":[{"id":"CWE-502","name":"Deserialization of Untrusted Data"}],"description":"The Keras Model.load_model method can be exploited to achieve arbitrary code execution, even with safe_mode=True.\n\nOne can create a specially crafted .keras model archive that, when loaded via Model.load_model, will trigger arbitrary code to be executed. This is achieved by crafting a special config.json (a file within the .keras archive) that will invoke keras.config.enable_unsafe_deserialization() to disable safe mode. Once safe mode is disable, one can use the Lambda layer feature of keras, which allows arbitrary Python code in the form of pickled code. Both can appear in the same archive. Simply the keras.config.enable_unsafe_deserialization() needs to appear first in the archive and the Lambda with arbitrary code needs to be second.","ecosystems":["PyPI"],"exploits":{"exploitdb":0,"first_seen":null,"github_poc":0,"max_poc_stars":null,"metasploit":0,"nuclei":0,"refs":[],"tier":"none_known"},"has_fix":true,"headline":"Arbitrary Code execution in Keras Safe Mode","id":"CVE-2025-9906","indexable":true,"kev":null,"kind":"cve","links":{"ecosystem":"PyPI","product":null,"products":[],"type":"deserialization","year":"2025"},"n_packages":3,"n_references":1,"next":{"headline":"Event-driven-ansible: event stream test mode exposes sensitive headers in aap eda","id":"CVE-2025-9907"},"packages":[{"ecosystem":"PyPI","fixed":[],"introduced":[],"last_affected":[],"package":"keras","purl":"pkg:pypi/keras","source":"PYSEC-2025-76"},{"ecosystem":"PyPI","fixed":["3.11.0"],"introduced":["0"],"last_affected":[],"package":"keras","purl":null,"source":"GHSA-36fq-jgmw-4r9c"},{"ecosystem":"PyPI","fixed":["3.11.0"],"introduced":["0"],"last_affected":[],"package":"keras","purl":"pkg:pypi/keras","source":"GHSA-36fq-jgmw-4r9c"}],"prev":{"headline":"Arbitary Code execution in Keras load_model()","id":"CVE-2025-9905"},"products":["Keras"],"published":"2025-09-19","related":[{"cvss_score":9.8,"headline":"Arbitrary code injection vulnerability in Keras framework \u003c 2.13","id":"CVE-2024-3660","kev_date_added":null,"published":"2024-04-16"},{"cvss_score":8.8,"headline":"Deserialization of Untrusted Data in keras-team/keras","id":"CVE-2026-12481","kev_date_added":null,"published":"2026-07-03"},{"cvss_score":8.8,"headline":"Safe Mode Bypass in keras-team/keras","id":"CVE-2026-1462","kev_date_added":null,"published":"2026-04-13"},{"cvss_score":8.6,"headline":"Keras safe_mode bypass allows arbitrary code execution when loading a malicious model.","id":"CVE-2025-8747","kev_date_added":null,"published":"2025-08-11"},{"cvss_score":8.1,"headline":"Path Traversal in keras-team/keras","id":"CVE-2026-11816","kev_date_added":null,"published":"2026-06-11"},{"cvss_score":7.3,"headline":"Arbitary Code execution in Keras load_model()","id":"CVE-2025-9905","kev_date_added":null,"published":"2025-09-19"}],"sources":[{"class":"attribution","licence":"CVE Terms of Use (MITRE)","name":"CVE List","prefix":"CVE"},{"class":"attribution","licence":"CC BY 4.0","name":"GitHub Advisory Database","prefix":"GHSA"},{"class":"attribution","licence":"CC BY 4.0","name":"PyPA Advisory Database","prefix":"PYSEC"},{"class":"attribution","licence":"US public information, NVD API terms","name":"NVD"},{"class":"attribution","licence":"CWE Terms of Use","name":"CWE (MITRE)"},{"class":"attribution","licence":"FIRST CVSS licence","name":"CVSS (FIRST)"}],"ssvc":{"automatable":"no","exploitation":"none","technical_impact":"total"},"state":"PUBLISHED","updated":"2026-02-26","vendors":["Keras-team"]},"summary":"Attackers can execute arbitrary code by loading a malicious Keras model archive that disables safe mode and exploits Lambda layers.","url":"https://insights.sbomify.com/CVE-2025-9906/"}